Open traverse, RTK, Pipe line

If The Control Was Not Too Good To Start With

What are you basing your work on now?

Paul in PA

> I am wondering what is the best way to set control with the RTK if at all.

Are we thinking that the RTK may be not be the best way to do this if at all?

In recent experience with RTK units it seems that .200' is about the best you will get without observing points with a base or closing a traverse angle and distance. RTK seems to only get within the ballpark of .200 if that is what you set your tolerance to, however even with a a tolerance set to .05 the collector will want to store it within a tenth due to whatever CQ you might have at your current location. RTK is great but not the gospel, just a tool.

>
> I am wondering what is the best way to set control with the RTK if at all. I have seen few posts about shooting points multiple times throughout the day initializing each time and then average them to create one point. I was also thinking of doing a resection using the averaged points.
>
> Thanks for any help

In my experience control can be set accurately with RTK, no problem at all. I usually like to set the unit cooking for up to 30 minutes over a point, on a tripod, same as I would if i were doing static or rapid static. I also usually get good results when averaging 2 x 1 minute observations on the same point separated by 10 minutes or more.

As with any GNSS campaigns do your constellation checking before you go if you want the best results. Observe your control under favorable satellite conditions.

A presentation by Janssen, Haasdyk and McElroy in 2011 containing analysis of days of GNSS data indicates observations are sufficiently independent from each other after about 20 minutes. A look at their graphs indicate de-correlation occurs fairly rapidly, and averaged observations even 5 minutes apart could be interpreted as being mostly independent.

I'd be disappointed if I were encountering more than the occasional RTK control point that wasn't agreeing with my TPS observations within 10 mm.

Mr. Stoner,

How about a few Rapid-Static observations as a check on your RTK control points?

Dave

it sounds like you are new to this type of work and may not have the tools you need. I would suggest you need to have static GPS capabilities and a good least squares program to adjust your traverse. RTK is not the best tool for this job.

I would take it easy calling the others work off 1.5'. If you set 2 gps points, occupied them conventially, traversed 3000', you very well could have "missed" by a foot and a half. To illustrate, set two hubs 100' apart. Set a third hub 10' from the first .1' off line. Pull a string across the two ten feet apart towards the 100' hub and see what you miss it by. Now the hub at 100' is "off" way more than your real error in the line of a tenth. Catch my drift?

> In recent experience with RTK units it seems that .200' is about the best you will get without observing points with a base or closing a traverse angle and distance. RTK seems to only get within the ballpark of .200 if that is what you set your tolerance to, however even with a a tolerance set to .05 the collector will want to store it within a tenth due to whatever CQ you might have at your current location. RTK is great but not the gospel, just a tool.

Poppycock! RTK can and will deliver the same as static, but requires some extra procedures. Multiple observations throughout the day, averaged will get the answer to within 0.03'. A single shot won't cut it (well it may but you'd never know).

However, and I can't stress this enough, many people have a fallacy of Maslow's hammer when it comes to RTK. It's not always the right tool to use. It's like the difference in MIG and stick welding. There is no doubt that MIG welding is great, but many times doesn't achieve the same strength as the stick, and it's not because of the properties of the electrode as much as it is the time spent with the rod reviewing and feeling the weld while laying it in. MIG fills holes way faster though and is superior in many ways when speed or large gaps are present. They just don't look as good.

> > MIG fills holes way faster though and is superior in many ways when speed or large gaps are present. They just don't look as good.

In the hands of a true craftsman, they do. 🙂

If you started from two bad points it could be off at the end how do you know whats good or bad . RTK is a lot better that .20 . If your trying to compare your Gun Trav to gps over a long run it will not give you the results you looking for , the key is to process the info in Least squares and then you will see the true error.

I find RTK will spot trav errors on big projects pretty quick . It is just a tool like a hammer but the new guys cannot swing a hammer , they hit everything but the hub. Learn to use the hammer first.

This is a waterline not a super collider. My experience is RTK with multiple observations works just fine for route surveys. I agree that you should not be seeing those kinds of discrepancies. I would suggest that you set intervisible control pairs with RTK every mile or so to tie into conventionally. The real question is what does the project need? Most likely, the contractor is more interested in the waterline's location relative to the edge of a road, a property line or other utilities than where it sits on the face of the earth.

A static point about every mile to check into with RTK is what I would do. The original control point of the other guy where probably sloppy RTK work.

> I am working on a water main pipe line project that has started a few months ago. The survey work has been done by a survey company, and not done very well. I now know the original traverse was set with GPS and not very well. I ran a traverse 3,000' between 3 original points, two were off 1.5' in different directions. There are only a few original points left due to construction. I am currently doing all the stake out, topo shots with a robot. Unfortunately I do not have time to run a traverse six miles and close it. I do have a RTK, and shot the last two points of my open traverse and I am .200'
>
> I am wondering what is the best way to set control with the RTK if at all. I have seen few posts about shooting points multiple times throughout the day initializing each time and then average them to create one point. I was also thinking of doing a resection using the averaged points.
>
> Thanks for any help

So, you are chasing "not very well" with "I really don't know what I'm doing". :'(

It sure sounds like a messy train wreck is about happen. Please hire someone who knows what they are doing before this blows up in your face.

> > > MIG fills holes way faster though and is superior in many ways when speed or large gaps are present. They just don't look as good.
>
> In the hands of a true craftsman, they do. 🙂

I are not one of those craftsmen. 🙂

Are you 100% positive that you're tying into the actual control points that you were provided? Have you asked for verification from your source, as to the coordinates?

It's pretty strange for GPS to be off 1.5' in 3000'. I'm wondering if you're either shooting into someone else's PK/MAG nails, or maybe whoever gave you the coordinates transposed a number or two.

And you're definitely on the same system, using US Survey feet, and not International?

I routinely use RTK to set control points, and when I check those same points with a total station the usual discrepancy is 0.01' horizontal. Never is it above 0.025 feet. Several years ago I took the unit to the standard baseline in Tucson, and saw very impressive horizontal and vertical results.

If you are seeing 0.2 feet something is wrong with your procedures or your instruments.

Check out this document from NGS.

You can do control with RTK. Static would be better, it is true. But if you can bear up with positions +/-0.03' or so, you can do control with RTK. You just can't do it willy-nilly.

1. Set base stations every couple of miles along your alignment. Set them in the wide open. Not necessarily right within the construction zone. But in the wide open. Vacant lots, City parks, school yards, etc. Observe time offset redundant vectors between them. 3-5 minutes. These sites are wide open and multipath free, so you are going to have good PDOP.

2. Collect data for OPUS static and get the extended solution from OPUS. That gives you error data on your OPUS positions. You will use these errors in your LS adjustment.

3. Set your project control within the construction zone, intervisible for TS use, from the base stations so established. Select points that are in the open well enough that you can get PDOPs under +/-3 with minimal multipath. Observe reduntantly (time offset), 60-90 second observations.

4. Any sites were you need control which yield PDOPs much in excess of 3 will have to be set with TS traverse.

5. LS adjust the vectors. You will have residuals in the 0.03' range. Maybe better.

If you do not have least squares I suppose you can still do control with RTK just by averaging redundant measurements and so on. LS is going to be a major improvement. I really recommend LS.

1

> If you are seeing 0.2 feet something is wrong with your procedures or your instruments.
When I first arrived in Oklahoma I was plunged into a turnpike survey that had control established with RTK. The field crew was dealing with errors between points routinely in the 0.2' range.

The control was all on the shoulder of the turnpike. Including the base station positions. Every time a semi went by several satellites would drop out. The base station was being set up on whatever point on the shoulder that was convenient that day. The crews had received no direction on how to set the control. They established it in the same way they were collecting topo. Set up where ever. Tie what ever. As long as you get "lock", you're good to go. The crew told me, "don't check into that one, unless your base is on that one, then you would get a good check". In other words, they had go-to check in points - for each base point - they knew would pass. Some days they would try checking into 2 or 3 point until they found one that was "good". Record that. Record that and go to town setting more "control". No redundant ties. No adjustment had been attempted by the PLS nominally in charge.

I know that this was not a unique circumstance. This is the way a lot of RTK work gets done.